Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Constructing a Computationally Secure Scheme Pseudorandomness Foundations of Cryptography Computer Science Department Wellesley College Fall 2016 Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Table of contents Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Pseudorandomness A pseudorandom string* • looks like a uniformly distributed string, as long as the entity that is ”looking” runs in polynomial time. Just as indistinguishability is • a computational relaxation of perfect secrecy, pseudorandomness is a computational relaxation of true randomness. *Technically no one string can be“pseudorandom”; pseudorandomness refers to adistribution of strings that indistinguishable from the uniform distribution. D Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Why study pseudorandomness? If a ciphertext looks random, then no adversary can learn any • information from it about the plaintext. This is the intuition behind the one-time pad. If a one-time pad used a pseudorandom string instead of a • truly random one, this should not make any di↵erence to a polynomial-time observer. The advantage is that a long pseudorandom string can be • generated from a relatively short random seed. Thus, a short key can be used to encrypt a long message. Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Pseudorandom Generators Definition 3.14. Let `( ) be a polynomial and let G be a · deterministic polynomial-time algorithm such that for any input s 0, 1 n, algorithm G outputs a string of length `(n). We say 2 { } that G is a pseudorandom generator if the following two conditions hold: 1. (Expansion:) For every n it holds that `(n) > n. 2. (Pseudorandomness:) For all probabilistic polynomial-time distinguishers D, there exists a negligible function negl such that: Pr[D(r) = 1] Pr[D(G(s)) = 1] negl(n), | − | where r is chosen uniformly at random from 0, 1 `(n),the { } seed s is chosen uniformly at random from 0, 1 n,andthe { } probabilities are taken over the random coins used by D and the choice of r and s. Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Not a pseudorandom generator To understand what makes for a secure pseudorandom generator, let’s examine a rather poor one. Example 3.15. Define G(s) to output s followed by n s with ⊕i=1 i expansion factor `(n)=n +1. Consider the distinguisher D that on input w, outputs 1 if and only if the final bit of w is equal to the XOR of all preceding bits. Since all strings output by G have this property, Pr[D(G(s)) = 1] = 1. On the other hand, if w is uniform, its final 1 bit is uniform and so Pr[D(w) = 1] = 2 . Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Pseudorandom generators are a far cry from random Consider G that doubles the length • of its inputs, that is, `(n)=2n. The uniform distribution over • 0, 1 2n chooses each of the 22n possible{ } strings with equal probability. In contrast, given an input of • length n, the number of di↵erent possible strings in G’s range is at most 2n. The probability that a given • random string is in the range of G n 2n n is at most 2 /2 =2− . Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Unlimited time works its magic* Consider the distinguisher D that works as follows: Upon • input string w, distinguisher D outputs 1 i↵ there exists an s 0, 1 n such that G(s)=w. 2 { } If w was generated by G,thenD outputs 1 with certainty. In • contrast, if w is uniformly distributed in 0, 1 2n then the { } n probability that there is an s with G(s)=w is at most 2− . We have • n Pr[D(r) = 1] Pr[D(G(s)) = 1] 1 2− , | − | ≥ − which is large. *Distinguisher D is our old friend Brute, and this type of attack is called brute force attack.Polynomial-timedistinguishersdonothavethetimetocarryout such an attack. Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Yes, Virginia ... We do not know how to • unequivocally prove the existence of pseudorandom generators. Nevertheless, we still believe in • them. We base this belief in part upon • the fact that they exist if one-way functions* do. *An assumption rather weaker than the existence of the tooth fairy if that is any comfort. Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Constructing a secure encryption scheme We construct a fixed-length • encryption scheme that has indistinguishable encryptions in the presence of an eavesdropper. Our scheme is very similar • to the one-time pad, except a pseudorandom string is used as the“pad” rather than a random string. Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions The encryption scheme Construction 3.17. Let G be a pseudorandom generator with expansion factor `. Define a private-key encryption scheme for messages of length ` as follows: Gen: On input 1n, choose k 0, 1 n uniformly at random • { } and output it as the key. Enc: On input a key k 0, 1 n and a message • 2 { } m 0, 1 `(n), output the ciphertext 2 { } c := G(k) m. ⊕ Dec: On input a key k 0, 1 n and a ciphertext • 2 { } c 0, 1 `(n), output the plaintext message 2 { } m := G(k) c. ⊕ Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Claim of indistinguishable encryptions in the presence of an eavesdropper* Theorem 3.18. Let G be a pseudorandom generator, then Construction 3.17 is a fixed-length private-key encryption scheme that has indistinguishable encryptions in the presence of an eavesdropper. *Note that our claim is not unconditional. Rather we reduce the security of the encryption scheme to the properties of G as a pseudorandom generator. Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Before the proof, a brief review is in order The experiment is defined for any private-key encryption scheme ⇧ =(Gen, Enc, Dec), any adversary ,andanyvaluen for the security parameter: A eav The eavesdropping indistinguishability experiment PrivK ,⇧(n) A 1. The adversary is given 1n, and outputs a pair of message m , m ofA the same length. 0 1 2 M 2. Akeyk is generated by running Gen(1n), and a random bit b 0, 1 is chosen. A challenge ciphertext c Enck (mb)is computed { } and given to . A 3. outputs a bit b0. A 4. The output of the experiment is defined to be 1 if b0 = b,and0 eav otherwise. We write PrivK ,⇧(n)=1 if the output is 1 and in this case we say that succeededA . A Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Adversarial indistinguishability Definition 3.8. An encryption scheme ⇧ =(Gen,Enc,Dec)had indistinguishable encryption in the presence of an eavesdropper if for all probabilistic polynomial-time adversaries there exisits a A negligible function negl such that eav 1 Pr[PrivK ,⇧(n) = 1] + negl(n), A 2 where the probability is taken over the random coins used by ,as A well as the random coins used by the experiment (for choosing the key, the random bit b, and any random coins used in the encryption process). Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Claim of indistinguishable encryptions in the presence of an eavesdropper* Theorem 3.18. Let G be a pseudorandom generator, then Construction 3.17 is a fixed-length private-key encryption scheme that has indistinguishable encryptions in the presence of an eavesdropper. Proof. Let ⇧ denote Construction 3.17. We show that if there exists a probabilistic polynomial-time adversary for which A Definition 3.8 does not hold, then we can construct a probabilistic polynomial-time algorithm that distinguishes the output G from a truly random one. Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions The adversary Let be a probabilistic polynomial-time adversary. We use to A A construct the following distinguisher D for the pseudorandom generator G. Distinguisher D: D is given as input w 0, 1 `(n). 2 { } 1. Run (1n) to obtain a pair of messages m , m 0, 1 `(n). A 0 1 2 { } 2. Choose a random bit b 0, 1 .Setc := w m . { } ⊕ b 3. Give c to and obtain output b .Output1ifb = b,and A 0 0 output 0 otherwise Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions A modified encryption scheme Define a modified encryption scheme ⇧ =(Gen, Enc, Dec) • that is exactly the one-time pad encryption scheme, except that we now incorporate a security parametere g thatg determinesg the length of the messages to be encrypted. Gen(1n) outputs a random key of length `(n), and the • encryption of a message m 0, 1 `(n) using the key 2 { } kg 0, 1 `(n) is the ciphertext c := k m. 2 { } ⊕ By perfect secrecy of the one-time pad, • 1 Pr PrivKeav (n)=1 = . ,⇧ A 2 h i e Introduction Pseudorandom generators Secure Encryption Scheme Multiple Encryptions Analysis of the distinguisher 1. If w is chosen uniformly from 0, 1 `(n), then the view of { } A when run as a sub-routine by D is distributed identically to the view of in experiment PrivKeav (n). Thus, ,⇧ A A e 1 Pr [D(w) = 1] = Pr PrivKeav (n)=1 = . ,⇧ w 0,1 `(n) A 2 { } h i e 2. If w is equal to G(k)fork 0, 1 n chosen uniformly at { } random, then the view of when run as a sub-routine by D is A distributed identically to the view of in experiment eav A PrivK ,⇧(n)and A eav Pr [D(G(k)) = 1] = Pr PrivK ,⇧(n)=1 .